<div dir="ltr">Hi M, The parameter in play here is called "upscale", take a look at the description of it in input parameters list. (Also for big systems one might want to keep in mind that conv_thr is an extensive quantity) cheers - emine<div><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Sat, Feb 7, 2015 at 3:11 PM, MKondrin <span dir="ltr"><<a href="mailto:mkondrin@hppi.troitsk.ru" target="_blank">mkondrin@hppi.troitsk.ru</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hello!<br>
<br>
I try to optimize quite large structure (several tens of atoms in the<br>
unit cell) but I can not quite get that is the criteria of convergence<br>
to proceed for the next optimization cycle. I use QE 5.01<br>
<br>
I have in the preamble of input file something like:<br>
<br>
&control<br>
calculation='relax',<br>
prefix='scf',<br>
tstress=.true.,<br>
tprnfor=.true.,<br>
pseudo_dir = '../pseudo',<br>
outdir='./tmp'<br>
/<br>
&system<br>
ibrav=0,<br>
celldm(1) = 12.60447,<br>
ntyp=2,<br>
nat=84,<br>
ecutwfc=70,<br>
ecutrho=800,<br>
occupations='fixed'<br>
/<br>
&electrons<br>
mixing_beta = 0.7,<br>
conv_thr = 1.0d-6<br>
/<br>
&ions<br>
/<br>
....<br>
<br>
I suppose that the next cycle of structure optimization should occur<br>
than the accuracy of self consistent calculations of about 1e-6 Ry.<br>
<br>
However in the output file I get this (this is the end of one scf cycle)<br>
<br>
iteration # 20 ecut= 70.00 Ry beta=0.70<br>
Davidson diagonalization with overlap<br>
ethr = 5.68E-10, avg # of iterations = 1.0<br>
<br>
total cpu time spent up to now is 230665.4 secs<br>
<br>
total energy = -823.97008715 Ry<br>
Harris-Foulkes estimate = -823.97008883 Ry<br>
estimated scf accuracy < 0.00000279 Ry<br>
<br>
iteration # 21 ecut= 70.00 Ry beta=0.70<br>
Davidson diagonalization with overlap<br>
ethr = 5.68E-10, avg # of iterations = 4.0<br>
<br>
total cpu time spent up to now is 233303.9 secs<br>
<br>
total energy = -823.97008869 Ry<br>
Harris-Foulkes estimate = -823.97008855 Ry<br>
estimated scf accuracy < 0.00000071 Ry<br>
<br>
iteration # 22 ecut= 70.00 Ry beta=0.70<br>
Davidson diagonalization with overlap<br>
ethr = 3.37E-10, avg # of iterations = 1.0<br>
<br>
total cpu time spent up to now is 236012.3 secs<br>
<br>
total energy = -823.97008898 Ry<br>
Harris-Foulkes estimate = -823.97008869 Ry<br>
estimated scf accuracy < 0.00000092 Ry<br>
<br>
iteration # 23 ecut= 70.00 Ry beta=0.70<br>
Davidson diagonalization with overlap<br>
ethr = 3.37E-10, avg # of iterations = 3.5<br>
<br>
total cpu time spent up to now is 238645.5 secs<br>
<br>
total energy = -823.97008880 Ry<br>
Harris-Foulkes estimate = -823.97008904 Ry<br>
estimated scf accuracy < 0.00000178 Ry<br>
<br>
iteration # 24 ecut= 70.00 Ry beta=0.70<br>
Davidson diagonalization with overlap<br>
ethr = 3.37E-10, avg # of iterations = 1.0<br>
<br>
total cpu time spent up to now is 241385.7 secs<br>
<br>
total energy = -823.97008829 Ry<br>
Harris-Foulkes estimate = -823.97008881 Ry<br>
estimated scf accuracy < 0.00000147 Ry<br>
<br>
iteration # 25 ecut= 70.00 Ry beta=0.70<br>
Davidson diagonalization with overlap<br>
ethr = 3.37E-10, avg # of iterations = 3.9<br>
<br>
total cpu time spent up to now is 244038.6 secs<br>
<br>
total energy = -823.97008797 Ry<br>
Harris-Foulkes estimate = -823.97008835 Ry<br>
estimated scf accuracy < 0.00000057 Ry<br>
<br>
iteration # 26 ecut= 70.00 Ry beta=0.70<br>
Davidson diagonalization with overlap<br>
ethr = 2.72E-10, avg # of iterations = 4.0<br>
<br>
total cpu time spent up to now is 246730.6 secs<br>
<br>
total energy = -823.97008832 Ry<br>
Harris-Foulkes estimate = -823.97008847 Ry<br>
estimated scf accuracy < 0.00000043 Ry<br>
<br>
iteration # 27 ecut= 70.00 Ry beta=0.70<br>
Davidson diagonalization with overlap<br>
ethr = 2.03E-10, avg # of iterations = 1.0<br>
<br>
total cpu time spent up to now is 249273.6 secs<br>
<br>
total energy = -823.97008831 Ry<br>
Harris-Foulkes estimate = -823.97008833 Ry<br>
estimated scf accuracy < 0.00000021 Ry<br>
<br>
iteration # 28 ecut= 70.00 Ry beta=0.70<br>
Davidson diagonalization with overlap<br>
ethr = 9.87E-11, avg # of iterations = 1.1<br>
<br>
total cpu time spent up to now is 252028.1 secs<br>
<br>
total energy = -823.97008826 Ry<br>
Harris-Foulkes estimate = -823.97008832 Ry<br>
estimated scf accuracy < 0.00000017 Ry<br>
<br>
iteration # 29 ecut= 70.00 Ry beta=0.70<br>
Davidson diagonalization with overlap<br>
ethr = 8.33E-11, avg # of iterations = 6.2<br>
<br>
total cpu time spent up to now is 254596.4 secs<br>
<br>
End of self-consistent calculation<br>
<br>
<br>
Aren't the cycles after #25 superfluous, because energy difference<br>
between two successive steps and estimated accuracy is less than 1e-6<br>
Ry? What is the control parameter that help me to reduce the number of<br>
steps ?<br>
<br>
Thank you in advance!<br>
<br>
M. V. Kondrin (High Pressure Physics Institute RAS, Troitsk, Russia)<br>
<br>
_______________________________________________<br>
Pw_forum mailing list<br>
<a href="mailto:Pw_forum@pwscf.org">Pw_forum@pwscf.org</a><br>
<a href="http://pwscf.org/mailman/listinfo/pw_forum" target="_blank">http://pwscf.org/mailman/listinfo/pw_forum</a><br>
</blockquote></div><br></div>